Assesses the retrieval effectiveness of automatically constructed interdocument hypertext links in information retrieval (IR). Describes experiments using statistical and probabilistic techniques that were designed to obtain evidence concerning the usefulness of querying and browsing automatically constructed IR hypertexts. Results indicate a…

Several norms for how people should assess a question's usefulness have been proposed, notably Bayesian diagnosticity, information gain (mutual information), Kullback-Liebler distance, probability gain (error minimization), and impact (absolute change). Several probabilistic models of previous experiments on categorization, covariation assessment,…

After a brief review of organization classification, multivariate approaches toward the empirical identification of such classifications are discussed. Ten guidelines for conducting future multivariate classificatory studies are presented, supported by a critique of two published large-scale empirical studies. (Author)

Sensitivity data is defined as involving two response categories, with responses observed at different levels of some variable. The responses are taken to indicate sensitivity to the variable and may be labeled "positive" or "negative." The countback method offers confidence limits for the 50% point, the level of the variable…

The Fisher Exact Probability Test considers independent groups and dichotomous criterion in a 2 by 2 (fourfold) contingency table. Fisher's test requires a determination of the combined probability of the observed frequency distribution and of the more extreme (in the same direction) frequency distributions. Typically, Fisher's test is for a small…

To resolve a recent controversy between Klein and Cleary and Levy, a model for dichotomous congeneric items is presented which has mean errors of zero, dichotomous true scores that are uncorrelated with errors, and errors that are mutually uncorrelated. (Author)

In this paper a reliable method is found for approximating the value of the Bivariate Normal Correlation Coefficient, rho, given values of the joint probability and the normal deviates, h and k, or the related areas. This technique finds useful application in the computation of the tetrachoric correlation coefficient, r, when the underlying…

The use of Bayesian statistics as the basis of classical analysis of data is described. Bayesian analysis is a set of procedures for changing opinions about a given phenomenon based upon rational observation of a set of data. The Bayesian arrives at a set of prior beliefs regarding some states of nature; he observes data in a study and then…

In this paper, a new item analysis index, e, is derived as a function of difficulty and discrimination to represent item efficiency. Item discrimination in this paper is not independent of item difficulty, and it is demonstrated algebraically that the maximum discriminating power of an item may be determined from its difficulty. Item efficiency is…

The three-parameter logistic model by Birnbaum for the multiple-choice item in the latent trait theory is considered with respect to the item response information function and the unique maximum condition. (Editor/RK)

A probabilistic unfolding model is presented which makes it possible to scale stimuli and subjects simultaneously, without assumptions concerning the distribution of subject or of stimuli. The applicability of the model is demonstrated in areas of attitude measurement where the assumption of normality is inadequate. (Editor)

This paper presents a procedure for estimating J scale (latent continuum) probabilities given a set of I scales (preference judgements). (Editor/RK)

Although a great deal of attention has been devoted over a period of years to the estimation of reliability from item statistics, there are still gaps in the mathematical derivation of the Kuder-Richardson results. The main purpose of this paper is to fill some of these gaps, using language consistent with modern probability theory. (Author)